How to Connect Shopify to Claude: Official Connector vs. Windsor.ai (3-Step Setup)

To connect Shopify to Claude, you can use two connectors from Claude’s directory that do different jobs. Shopify’s own connector, listed since April 2026, runs the store: it can build a new store, manage inventory, create discount codes, review orders and customers, and show store analytics. The Windsor.ai connector analyzes the business: it gives Claude your Shopify orders, customers, products and refunds next to your ad platforms, GA4 and CRM.

For the Windsor.ai route, connect your store at onboard.windsor.ai/app/shopify with the Client ID and Client Secret of an app you create in the Shopify Dev Dashboard, then add the Windsor.ai connector in Claude. Start a new chat and ask about margins, repeat buyers or which ad campaigns bring revenue.

Many stores use both: Shopify’s connector to change things in the store, and Windsor.ai when the question needs data from outside Shopify. Windsor.ai reads Shopify and does not change your store.

🚀 Connect Shopify to Claude with Windsor.ai. Free forever plan, no credit card → onboard.windsor.ai/app/shopify.

3 steps to automatically sync Shopify data to Claude via Windsor MCP

Setup is a one-time connection in Windsor.ai and one click in Claude.

📖 Follow our quick step-by-step guide to connect your Shopify data to Claude.

Prerequisites for Shopify to Claude Integration

Before you begin, ensure you have the following ready:

• A Windsor.ai account: You need an active account to integrate the Shopify data.
• A Shopify app for your store: You create it in the Shopify Dev Dashboard, and it gives you the Client ID and Client Secret that Windsor.ai asks for. The Shopify Configuration Guide shows each step.
• Claude.ai account: You need an active Claude account, which serves as your AI analytical environment.

Step 1: Connect your Shopify store to Windsor.ai

1. Log in to onboard.windsor.ai.
2. Select Shopify as your data source.
3. Click Connect Shopify Store and enter your shop name, Client ID and Client Secret. To get the Client ID and Client Secret, create an app for your store in the Shopify Dev Dashboard as shown in the Shopify Configuration Guide.

Windsor.ai then reads your store through the Shopify API, with fields from customer_total_spent to order_net_sales.

Step 2: Activate the Windsor.ai connector in Claude

In Claude, open Customize, then Connectors, search for Windsor and click Add. Sign in to your Windsor.ai account when prompted. You can also open the Windsor.ai listing in the Claude directory and click Connect.

Optionally, you can set permissions to “Always allow” to enable Windsor MCP-powered queries across all your chats.

Step 3: Start querying your Shopify data

Once connected, you no longer need to look at static analytical dashboards. Just speak to Claude in natural language to extract instant AI insights:

Here are prompt samples you can use:

• “What is my average Customer Lifetime Value (LTV) for customers who used a {WELCOME10} discount code vs. those who didn’t?”
• “Which products are my ‘hidden gems’: low sales volume but extremely high repeat purchase rates?”
• “Analyze my refunds (order_refunds_subtotal) by product and tell me which SKUs are losing the most revenue to returns.”

Dig deeper and ask Claude to generate a visual summary based on the text output.

Official Shopify Claude connector vs. Windsor.ai vs. other methods

• Shopify’s own connector (in Claude’s directory): best for running the store by chat. Its listing says it can build a new store with Shopify’s AI store builder, manage inventory across locations, create discount codes, review orders and customer details, and pull store analytics. It reads and changes your store.
• Windsor.ai connector (in Claude’s directory): best for questions that need more than store data, such as revenue by ad campaign after refunds, or repeat purchase rate by traffic source. It gives Claude 350+ sources next to Shopify and only reads Shopify; the connectors that can make changes are listed in the write actions documentation.
• Shopify’s AI Toolkit with Claude Code: developer tooling from shopify.dev for building apps and themes and working with Shopify’s APIs from the terminal.
• A CSV export from the Shopify admin: for a one-off check. Claude analyzes that file only and cannot refresh it.

Claude can use several connectors in the same chat, so Shopify’s connector and Windsor.ai can run side by side.

Can Claude build a Shopify store for me?

Yes, through Shopify’s own connector. Its directory listing describes turning a business idea into a store with Shopify’s AI store builder: you describe what you sell, and the store and products are created from the chat. Windsor.ai does not build or edit stores; it analyzes the data once the store is selling.

Can I use Claude Design for my Shopify store?

Claude Design, from Anthropic Labs since April 2026, makes prototypes, slides and one-pagers from a chat. Anthropic does not describe it as a store builder. Some Shopify page-builder apps offer their own Claude connectors to turn a design into a store page. Windsor.ai is not part of that workflow.

Should I use the official Shopify connector or Windsor.ai?

Use Shopify’s connector when you want to change something in the store. Use Windsor.ai when the question needs data that Shopify does not hold, such as ad spend from Meta Ads or Google Ads, site behaviour from GA4, or deals from your CRM, next to Shopify revenue.

Things you can do with your Shopify data in Claude

Use cases that work with Shopify data in Claude:

• Sales & profitability analysis: Claude can analyze order_gross_sales, order_total_discounts and order_net_sales to show net sales after discounts and returns, month by month.
• Deep customer insights: Use fields like customer_total_spent, customer_orders_count, and order_total_price to let Claude identify your highest-value customers and what they have in common.
• New vs. returning customer analysis: Segment orders by order_new_or_returning_customer and order_created_at to compare revenue, order counts, and repeat purchase patterns. Identify which channels and products drive first-time vs. repeat purchases.
• Churn prevention: Identify customers who haven’t purchased in 60 days but have a high customer_orders_count. Then ask Claude to draft a personalized “Win-back” email.
• Inventory velocity: Ask Claude to compare line_item__quantity sales trends with product_variant_inventory_available_quantity to estimate which line_item__sku will run out in the next 14 days.
• B2B/wholesale discovery: Use customer_tags to separate retail from wholesale buyers and compare their average order value from order_total_price and order_count.
• Marketing ROI by channel: Connect your sales back to the source. Using order_customer_last_visit_source and order_customer_last_visit_utm_source, Claude can tell you if your latest influencer campaign or TikTok ad brought repeat buyers or one-time buyers.

Shopify analytics prompt ideas for Claude

But the questions that drive growth are harder:

• Which customers generate the most profit?
• Which products tie up money without selling?
• Which marketing channels bring high-value buyers?
• Which SKUs will run out of stock next week?

These questions usually require exporting Shopify data into spreadsheets or BI tools.

Once your Shopify data is connected to Claude through Windsor MCP, you can ask these questions directly and get structured AI-driven answers.

Below are some common e-commerce problems and the prompts you can use to solve them.

1. Are you making money on each order?

Revenue in Shopify does not equal profit.

Discounts, taxes, shipping costs, and refunds can significantly reduce margins. Many stores rely on rough estimates instead of calculating true net revenue.

Claude can analyze order-level data such as:

• order_gross_sales
• order_total_discounts
• order_net_sales

to estimate real profitability.

Example prompt:

Compare order_gross_sales, order_total_discounts, and order_net_sales by month.
Which month had the highest net margin percentage, and how much revenue did discounts cost me?

This shows which months kept the most revenue after discounts.

2. Which customers are your most valuable?

Not all customers are equal.

Many stores focus on acquisition while ignoring the customers who generate the majority of lifetime value.

Using fields such as:

• customer_total_spent
• customer_orders_count
• order_total_price

Claude can identify high-value customers and purchasing patterns.

Example prompt:

Identify the top 10% of customers based on customer_total_spent.
What is their average customer_orders_count and average order value compared to the rest of the customer base?

This helps you build ideal customer profiles (ICPs) and optimize acquisition strategies.

3. Are you losing repeat customers without noticing?

Customer churn is one of the biggest hidden revenue leaks for e-commerce stores.

Customers who previously ordered multiple times but suddenly stopped buying are often the easiest revenue to recover.

Claude can detect churn risks by analyzing:

• customer_orders_count
• order_created_at

Example prompt:

Find customers with a customer_orders_count greater than 3 whose most recent order_created_at is more than 60 days ago.
Draft a personalized win-back email referencing the last line_item__title they purchased.

This allows you to turn churn analysis directly into retention campaigns.

4. Which products will run out soon, based on how fast they sell?

Stockouts can cost stores significant revenue, especially for best-selling products.

Many Shopify operators manually track top-selling products and declining inventory (“downsells”) to forecast inventory levels.

By analyzing fields such as:

• line_item__quantity
• line_item__sku
• order_created_at

Claude can estimate how fast each product sells.

Example prompt:

Identify my top 10 selling SKUs over the last 30 days.

For each SKU:
- Calculate daily sales velocity
- Estimate days until stockout (based on recent trends)

Highlight:
- Products at risk of running out within 21 days
- Slow-moving products with declining demand

Provide a prioritized reorder list.

This helps operators prevent lost revenue from unexpected stockouts.

5. Which marketing channels bring high-value customers?

Not all traffic sources produce the same type of customer.

Some channels generate one-time bargain hunters, while others produce long-term repeat buyers.

Using fields like:

• order_customer_last_visit_source
• order_customer_last_visit_utm_source
• customer_total_spent

Claude can analyze which channels drive the highest long-term value.

Example prompt:

Compare order_customer_last_visit_source and order_customer_last_visit_utm_source values.
Which traffic source produces customers with the highest customer_total_spent over time?

This helps marketers optimize acquisition toward channels that drive repeat purchases and LTV.

6. Which products are dragging your margins down?

Many Shopify stores carry products that:

• sell rarely
• tie up inventory
• generate little revenue

Claude can analyze product sales activity and identify inventory that is no longer moving.

Example prompt:

Identify active SKUs with zero sales in the last 60 days.
Calculate their total inventory value from product_inventory_value and suggest a discount or bundle to clear the stock.

This helps operators free up capital and warehouse space.

7. How many new vs. returning customers do you have?

Understanding which orders come from first-time buyers versus repeat customers is the basis for measuring growth, retention, and the effectiveness of marketing campaigns. Claude can automatically segment your Shopify orders using fields like:

• order_new_or_returning_customer
• order_created_at
• customer_total_spent

Example prompt to compare the amount and revenue:

For the last 90 days, calculate the total number of orders and revenue from new customers vs. returning customers and their average order value.

Example prompt to define the top-performing channels:

Identify which marketing channels bring the highest proportion of new customers vs. returning customers.

Example prompt to discover the best-selling products:

Analyze repeat purchase rates by product category and highlight which items drive loyalty among returning buyers.

This explains whether growth is driven by acquisition or retention and helps you optimize campaigns for repeat buyers and improve customer lifetime value.

Bonus: Which advertising campaigns generate the highest true ROAS?

Most e-commerce teams run paid ads across Facebook Ads, Google Ads, TikTok Ads, etc., but struggle to analyze Shopify revenue and ad performance in separate dashboards.

Native platform analytics report conversions and estimated ROAS, but they don’t always reflect actual store revenue.

To calculate true campaign performance, you need to combine:

Shopify data:

• order_total_price
• order_total_discounts
• order_customer_last_visit_utm_campaign

Advertising data (f.e, Meta Ads):

• campaign_name
• spend
• clicks

Using UTM campaign parameters, Claude can match Shopify orders to Facebook Ads campaigns and calculate real ROAS based on actual store revenue.

Example prompt:

Match Shopify orders with Facebook Ads campaigns using order_customer_last_visit_utm_campaign.
For the last 30 days, calculate for each campaign:
- total ad spend
- total Shopify revenue (order_total_price)
- ROAS
Rank campaigns by ROAS and highlight the best-performing ones.

This allows marketers to quickly identify:

• Which campaigns generate the highest ROAS
• Which campaigns should receive more budget
• Which campaigns are underperforming

Instead of relying on isolated dashboards, Claude can analyze both revenue and ad spend together, giving you a clearer view of real marketing performance.

Shopify data that you can stream into Claude with Windsor.ai

Windsor.ai returns 435 Shopify data fields.

Here are the primary data categories you can sync:

Sales & transactional data

This category includes everything related to the financial health of your orders. It allows Claude to audit your actual take-home pay after costs.

• Key fields: order_total_price, order_net_sales, order_gross_sales, order_total_discounts, and order_currency.

Customer intelligence

These fields help Claude perform cohort analysis and identify your most valuable buyers.

• Key fields: customer_total_spent, customer_orders_count, and order_new_or_returning_customer.

Product & inventory performance

This data is essential for Claude to act as an inventory planner, identifying what’s moving and what’s stuck.

• Key fields: line_item__sku, line_item__quantity, line_item__title, and product_status.

Marketing and channel performance

This category connects your sales back to the source, helping Claude determine which channels are profitable.

• Key fields: order_customer_last_visit_source, order_customer_last_visit_utm_source, order_customer_last_visit_utm_campaign, and order_sales_channel.

The manual way to send Shopify data to Claude

If you only need a one-time snapshot of your store, you can follow these steps to connect Shopify to Claude:

1. Export orders: Navigate to Shopify Admin -> Orders -> Export. Select “All Orders” and choose “CSV for Excel, Numbers, or other spreadsheet programs.”
2. Export customers: Go to Customers -> Export to get your buyer history.
3. Data cleaning: Open the files in Excel to remove sensitive PII (like full addresses) and ensure date formats are consistent.
4. Upload to Claude: Open Claude.ai, start a new chat, and use the attachment icon to upload your cleaned CSVs.
5. Prompt: Use a prompt like: “I have attached my Shopify order and customer data. Please analyze these files to identify which product categories have the highest repeat purchase rate.”

Limits of a manual export:

• Orders, customers and products export as separate files, so Claude has to join them, and any mismatch in IDs or dates shows up in the answer.
• An export is a snapshot: orders and refunds after the export are missing.

Limits to know before you start

• Shopify gives apps the last 60 days of orders by default. Older orders need the read_all_orders scope on your app (Shopify’s access scopes); the Shopify Configuration Guide shows how to request it.
• Windsor.ai’s free forever plan (no credit card) covers 1 data source and 1 account with the last 30 days of data. Longer history, more stores or Shopify next to ad platforms needs a paid plan: Basic is $19/month for 3 sources, Standard is $99/month for 7 (annual billing, see pricing).
• Shopify limits how many API requests each app can make, and each app has its own limit (Shopify’s API limits). Windsor.ai’s requests go through the app you created for Windsor.ai.
• Windsor.ai reads Shopify and never changes products, orders or inventory. Payment card details are not part of the data.

Why Claude’s numbers can differ from Shopify admin

When a total in Claude does not match a Shopify report, check these four things first:

• Time zone: Shopify’s reports use the store’s time zone. If the answer groups orders by a different time zone, orders placed near midnight fall on a different day.
• Which total: Shopify’s Total sales adds shipping, taxes and duties to net sales. Compare net sales with order_net_sales, and the amount customers paid with order_total_price.
• Refund dates: Shopify’s sales reports count a refund on the day it was made. A total grouped by order date counts it in the month of the order.
• Filters: use the same date range and sales channel (order_sales_channel) in both.

Sending Shopify data to Google Sheets or BigQuery

Claude reads Shopify through the connector, so asking questions in a chat needs no export. For a stored copy that other tools read, such as a Data Studio dashboard or a warehouse table, Windsor.ai can write the same data to Google Sheets, BigQuery, Snowflake and other destinations on a schedule you set. Destinations are included on every Windsor.ai plan.

Using Shopify data in Claude Code

Claude Code can use the same Windsor.ai connection. Install the Windsor.ai plugin with /plugin install windsor-ai@claude-plugin-directory, then sign in to Windsor.ai when prompted.

Pull last month's Shopify orders with order_net_sales, order_discount_code and order_customer_last_visit_utm_campaign.
Save them to shopify_orders.csv and add a column with revenue per UTM campaign.

Conclusion

Shopify’s own connector runs the store from Claude. The Windsor.ai connector puts Shopify revenue, customers and products next to your ad and analytics data, so Claude can answer margin, retention and channel questions. Start with the margin prompt, then the ROAS prompt if you run ads.

🚀 Connect Shopify to Claude. Free forever plan, no credit card → onboard.windsor.ai/app/shopify.